Quasi-Experimental Evaluation Without Regression Analysis

In Brief Evaluators of public health programs in field settings cannot always randomize subjects into experimental or control groups. By default, they may choose to employ the weakest study design available: the pretest, posttest approach without a comparison group. This essay argues that natural experiments involving comparison groups are within reach of public health program managers. Methods for analyzing natural experiments are discussed. The purpose of this commentary is to argue that natural experiments abound, that they offer better validity than simple pre-post designs, that their use should be taught in public health graduate programs, and that they should be employed more frequently in public health practice.